Project Overview


Australian Government Department of Agriculture as part of its Rural R&D for Profit program, Meat and Livestock Australia, Dairy Australia and Australian Wool Innovations Ltd.

Primary Research Advisor:

Dr Richard Simpson, CSIRO.

Project Duration:

Four years (2016-2020)

Project Host(s):
  • Geoff and Bev Kowald – Hutton Road, Carrolup
  • Mark Paganoni – Fletcher Road, Broomehill     


  • Development of low-P pasture systems based on highly P-efficient serradellas (Ornithopus spp.), using a combination of farmer-participatory and traditional research.

  • Identification and then promotion of the most P- efficient cultivars of subterranean clover (Trifolium subterranean).

  • Development of the knowledge and genetic/agronomic protocols necessary for development of subterranean clover cultivars with P- efficient root traits analogous to those of serradella species.

  • Updates to the industry decision support tool (“Five Easy Steps” soil-P management tool, tools/producer-case-studies/five-steps-to-better- phosphorus-use/) to include the principles and management that underpin the achievement of highly-productive, P-efficient pasture systems. This tool applies to pastures grown on acid soils across southern Australia (~35 m/ha) and is already in use in many farming districts.

Preliminary Findings:

Both trial sites performed well in 2018, however no statistical differences were found between the P treatments for the dry matter kg/ha. At the start of the project site selection was made on P levels from soil core samples taken at 0-10 cm depth. In the spring of 2017 soil core samples were taken at 0-10 cm and 0-60 cm depths. It was established that P was pooling in these soils at around 30-60 cm. This legacy P is likely utilised by the plants, therefore meeting all P requirements. It is unlikely that we will be able to establish the critical P levels from these sites due to the higher P levels at depth.

The finding of legacy P at depth in these soil types promotes the discussion for the need to change and expand soil sampling procedure and protocols to better understand and advise on fertiliser-P requirements.


Phosphorus is the primary nutrient input that drives legume growth and nitrogen fixation in pastures of southern Australia. It underpins high productivity and profitability. However, P- fertiliser costs have doubled since 2000 and a tightening in future supply seems inevitable. Both factors are expected to lead to a steady increase in the fertiliser cost of production.

It is estimated from previous research (Simpson et al. 2014) that it will be feasible to develop pasture systems that require ~30% less P fertiliser annually than current sub clover-based pasture systems. The research suggests this should be possible via pasture legumes, such as serradellas that yield better in low P soils and have lower ‘critical’ soil-test P requirements for maximum yield than sub clover and annual medics.

The project builds on research that has shown that P accumulation in a grazing system soil is significantly reduced when it is managed at a lower soil test P concentration (Simpson et al. 2015), and on work that has identified high- yielding pasture legumes with lower P requirements (Sandral et al. 2015). Together, these innovations mean that low-P grazing systems are now within reach. Pasture systems operated at lower soil test P concentrations should reduce the risks of P loss from agriculture to the natural environment as well as reducing input costs.

Trial Details

Two sites are paddock scale trials located in Katanning and Broomehill, the experiment will run for three years from 2017 to 2020. The main objective of these experiments is to test the hypothesis that serradella pastures can yield well at soil test P levels that are considerably lower than those expected to be necessary for high yielding sub clover pastures.

Initial soil tests 0 – 10 cm were collected from each site to determine phosphorous levels. The sites “ideally” required low phosphorous levels (approx 12 mg/kg Colwell P or 6 mg/kg Olsen P) with Phosphorus Buffering Index values (PBI) >50.

Each trial had three treatments of a low, medium, and high phosphorous rate. The trial is a randomised block design, each plot 4 m by 10 m, with 1 m buffers. Triple super was applied at rates of 0 kg/ha, 21.5 kg/ha and 43 kg/ha (21% P). To eliminate other deficiencies at the site, which may have negative implications on the trial, 100 kg/ha of KCI, 100 kg/ha Gypsum, and micronutrients, where applied to the entire site as per Bolland and Paynter (1992). The treatments were applied to each plot by hand ensuring an even application of the phosphorous treatments. The treatments are applied in March, or the break of season of each year.

Figure 1a. Trial site layout for Katanning


Figure 1b. Trial site layout for Broomehill

Assessments will be carried out each year over the three-year trial. The assessments include spring soil sampling, where a sample of 0-10 cm will be collected in spring while the soil is moist to monitor soil P levels. Seedling counts will be carried out approximately one month after the break of season.

Frequency counts will be collected in late winter before more than 1500 kg DM/ha accumulates. Dry Matter Measurements and botanical composition will be collected October to November; four to five weeks after site lock-up.

Botanical assessments and dry matter cuts were collected from the Katanning site on the 4th October 2018. The botanical composition was measured as per the project protocol. This was carried out by estimating within 10 quadrats per plot, walking diagonally across the plot and recording every two steps using a 0.1 m2 quadrat. This was repeated in each plot. Data were recorded and entered in the provided excel spreadsheet. The botanical compositions were made up of serradella, sub clover, broadleaf species, and annual grasses.

Dry matter cuts were also collected to determine pasture production of each plot and treatment. After the lock up period, 10 by 0.1 m2 quadrat cuts were collected per plot. The cuts were taken using shears which cut 1 cm above ground level. The samples were bulked into a single sample per plot. The samples were taken to the Department of Primary Industries and Regional Development to be weighed fresh and then put in drying ovens to measure dry weight. The samples were dried from 4th October afternoon to 8th October 2018. Weights were recorded and entered into the excel spreadsheet for analysis.

Soil samples were collected from the site in early December, using a soil auger. As per the agreed changes to the soil sampling protocol, each plot has 15 by 0 – 10 cm collected using a soil pogo stick, then bulked into one sample. Another sample of 15 by 10 – 60 cm was collected using a soil auger, then bulked into one sample. This was repeated in each plot.

The soil samples where then packaged and sent to CSIRO lab in Canberra for analysis.

2018 Results

Botanical assessments and dry matter cuts were collected from both the Katanning and Broomehill sites in October. Dry matter cut measurements were also collected with both fresh and dried weights recorded. No significant findings were found from the 2018 trial site data, however the soil testing results have highlighted some potential problems that may arise at the initial stages in site selection of a project.

KATANNING - Demonstration Site One

The botanical composition observed over most plots, was dominated by the serradella species. Annual grasses and broadleaf weeds were abundant in each plot, as seen in Figure 2. Sub clover was found in only a few plots, this was likely due to the dry start to the season, and fits with the plant establishment count and frequency count taken early in the season. Perennials were not found at the site.

Figure 2. Botanical composition (%) for each plot at the Katanning site WA, 2018.

Figure 3. Katanning site, 2018. Image captured via drone.

The dried dry matter cuts which established pasture productivity in kg/ha showed the low phosphorous rate plots on average yielded 1596 kg/ha. The medium plots yielded 1556 kg/ha on average and the high P plots 1270 kg/ha on average. No phosphorous was applied in 2018 to the plots as there was sufficient P in the soil according to the 2017 spring soil tests. We are awaiting the results of the 2018 spring soil tests.

BROOMEHILL - Demonstration Site Two

The serradella was also the dominant species found in the botanical composition assessments at the Broomehill site (Figure 4). Broadleaf weed, which were comprised of mainly capeweed and erodium, were also in high numbers in each plot. Sub clover was only found in three plots in a small section of trial.

Figure 4. Botanical composition (%) for each plot at the Broomehill site WA, 2018.

Figure 5. Broomehill site, 2018. Image captured via drone.

The dry matter cuts found in the high phosphorous plots the yield was 2770 kg/ha on average. The medium P plots were 1853 kg/ha and low P plots 1946 kg/ha on average. As at the Katanning site, there was no phosphorous applied to the plots as there was a substantial amount of phosphorous in the soil sampled collected in spring 2017. We are awaiting the results of the 2018 spring soil tests.

Figure 6. Katanning monthly rainfall data for 2018.

Figure 6. Broomehill monthly rainfall data for 2018.


The 2018 season started out very dry, but the break of the season arrived in June, and we continued to get good rainfall events throughout the growing season with excellent finishing rains in September and October (Figures 6 and 7). The results of the data we collected in the 2018 season are yet to confirm the projects aims of the effectiveness of confirming in farm paddocks that maximum production of farm scale serradella- based pastures can occur at its expected low critical P requirement (less than 10 mg P/kg). A major issue we have come across through this project were in the initial stages while looking for suitable trial sites. Soil samples were collected at 0 – 10 cm and the sites were selected on those results. In 2017 the spring soil results showed from soil cores taken 0 – 60 cm there were luxurious level of P in the 30 – 60cm profile. Therefore, the plants were likely utilising this P. A positive to come out of this project has been the discussion between local agronomists and other researchers about the depth we should be collecting our initial soil samples from, as well as ongoing sampling.

The lack of sub clover and the dwindling number of serradella plants, particularly at the Katanning site, has prompted some discussion around whether or not to re-sow serradella and sub clover in 2019.

The Southern Dirt P4pasture trials performed well in the 2018 season compared to many in NSW and Victoria. We will await to find out the results of the soil tests and discuss how we will proceed in 2019. This milestone has been met in full.

In September 2018, Southern Dirt held its annual Spring Field Day event. One of the trials we stopped at was the Katanning P4Pastures site, where we had Brad Nutt present the project and his trial results. There were 50 farmers and industry professionals attending the trial tour, with excellent feedback given about the trial.

The trial continues throughout 2019 and into 2020.